Just wondering if you've tried it for real : ) I attempted to linearized with two thermistors, but didn't get great results : (
I didn't spend long enough to figure out why, but I suspected it could be because they were not thermally coupled, they were separated by about 5 mm. It may work better if they are thermally glued together maybe! I had tested with SMD thermistors (I think they were 0603), to measure air temperature, and maybe the low thermal mass was exacerbating the problem.
There are ready-made dual thermistors that are in a single package, and they provide great results, so that was the reason why I thought maybe the separated ones were not functioning well.
I really hope you get it to work! It would be an awesomely low-cost way to get excellent measurements for (say) ambient temperature, since thermistors can cost just pennies, and lots of room thermostats perhaps uses cheap silicon sensors (not the good ones) - at least the one I have at home is really bad.
The main problems I was getting in the limited time I was testing the dual thermistors, was that despite the environment getting warmer, I'd see the measurement sometimes drop, i.e. it was not monotonic, presumably because one thermistor was warming up before the other one, even just 5 mm apart, and for me the error was quite high too, although perhaps a better computer simulation could have picked an improved combination that could handle such differences better until both thermistors reach equilibrium. Bonding two thermistors could be neat too. One thing I wanted to experiment with was to bond a thermistor to a 50 ohm resistor, to make a cheap RF power meter. I bought some 0402 thermistors and 0402 50 ohm resistors to attempt that, but it's so fiddly : )
